Traffic image intelligent analysis method based on big data

Abstract

The invention relates to the technical field of traffic intelligent analysis, and discloses a traffic image intelligent analysis method based on big data, which comprises the following steps: determining a linear road and a starting point and an ending point thereof, obtaining traffic signal lamp information, dividing the road into a plurality of road sections according to traffic signal lamps on the road, recording states of different signal lamps, and determining a traffic image; acquiring a navigation route of a vehicle in each lane, determining a recommended vehicle, acquiring a traffic signal lamp of a current road section according to a driving road section of the recommended vehicle, analyzing and processing each lane image in the road section, and acquiring vehicle information of each road section through image analysis and processing; and analyzing and screening the vehicle information of each lane under different traffic lights, and recommending an optimal lane according to a screening result.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent traffic analysis, and in particular to a traffic image intelligent analysis method based on big data. Background Art

[0002] Car navigation is an assisted driving technology developed based on the global satellite positioning system GPS. As long as the driver enters the destination into the car navigation system, the system will automatically calculate the most suitable route based on the electronic map and remind the driver to follow the calculated route while the vehicle is driving.

[0003] Nowadays, the application of car navigation technology has greatly improved people's travel experience. It can not only help drivers plan their car's driving route, but also provide real-time traffic information to avoid congestion. In addition, with the development of Internet of Vehicles technology, car navigation systems can also communicate with other vehicles and traffic infrastructure to improve road safety.

[0004] However, with the rapid development of urbanization and the more convenient infrastructure, the number of private cars has increased significantly. With the increase in vehicles, traffic congestion has become more and more serious, and some bottlenecks have appeared on traffic arteries. The existing navigation technology can no longer effectively alleviate the traffic congestion. Moreover, the existing navigation technology can only recommend driving routes for vehicles, but cannot recommend specific driving lanes for vehicles in real time according to the traffic lights on the road. When special lanes appear, it is easy to cause traffic congestion, affecting the vehicle traffic efficiency of the entire road.

[0005] In summary, there is an urgent need for an intelligent traffic image analysis method based on big data to solve the above problems. Summary of the invention

[0006] The present invention provides a traffic image intelligent analysis method based on big data, which is used to promote the solution of the problems mentioned in the above background technology.

[0007] The present invention provides the following technical solution: a traffic image intelligent analysis method based on big data, comprising:

[0008] Determine the straight road type;

[0009] According to the straight-ahead direction in the straight road, the two ends of the straight road are recorded as the starting point and the ending point;

[0010] Get all traffic lights within a linear road;

[0011] Divide a straight road into multiple road segments according to traffic lights;

[0012] Get traffic light change information;

[0013] Traffic light information includes left turn red light and straight ahead red light, left turn red light and straight ahead green light, left turn green light and straight ahead red light;

[0014] Determine the recommended vehicles and enter the target set;

[0015] The lane recommendation strategy is performed on the elements in the target set in turn. The lane recommendation strategy is as follows:

[0016] Get the lane information of the current road segment, which includes multiple straight lanes and a left turn plus straight lane;

[0017] Obtain the traffic light information of the road section where the target vehicle is currently traveling;

[0018] According to the traffic light information of the road section where the target vehicle is currently traveling, the image of each lane under the current road section is obtained and analyzed and processed;

[0019] According to the image analysis processing result of each lane in the current road section, the vehicle information of each lane in the current road section is determined;

[0020] The best lane that matches the recommended vehicle navigation route is determined through the vehicle information and traffic light information of each lane of the current road segment.

[0021] Optionally, determining the recommended vehicles and entering the target vehicle set includes:

[0022] Sort all road segments from nearest to farthest from the starting point;

[0023] The first ranked road segment is taken as the target road segment;

[0024] Obtain the navigation routes of all vehicles in the target road segment;

[0025] The vehicles whose navigation routes are to pass the current traffic light and go straight are taken as recommended vehicles and are entered into the target vehicle set;

[0026] Sort all elements in the target vehicle set according to the order in which they pass the traffic lights of the current road segment;

[0027] Select the elements in the target vehicle set as the target vehicle in sequence.

[0028] Optionally, the acquiring, based on the traffic light information of the road segment on which the target vehicle is currently traveling, an image of each lane of the current road segment and performing analysis and processing includes:

[0029] Get the traffic light of the current road section of the target vehicle;

[0030] If the traffic lights of the current road section of the target vehicle are a left-turn red light and a straight-through red light, then the straight-through image and the left-turn plus straight-through image of the current road section are obtained;

[0031] The through lane image and the left turn plus through lane image are respectively identified by image recognition technology;

[0032] According to the image recognition results, the number of vehicles in each straight lane and the number of vehicles turning left and going straight are obtained respectively;

[0033] All lanes are entered into the first vehicle judgment set in ascending order according to the number of vehicles in each lane.

[0034] Optionally, determining the best lane that meets the recommended vehicle navigation route through the vehicle information and traffic light information of each lane of the current road segment includes:

[0035] Select the elements in the first vehicle judgment set in sequence and record them as pre-selected lanes;

[0036] If the pre-selected lane is a straight road, the pre-selected lane is recommended as the best road and the lane recommendation strategy ends;

[0037] If the pre-selected lane is a left-turn and straight lane, the navigation routes of all vehicles in the left-turn and straight lane are obtained;

[0038] Mark all vehicles whose navigation routes are left turns and enter the first left turn set;

[0039] If the number of elements in the left-turn set is zero, then the left turn plus the straight road is recommended as the best road, and the lane recommendation strategy ends;

[0040] If the elements in the left-turn set are not zero, the left-turn plus straight lane in the first vehicle judgment set is eliminated, and the elements in the first vehicle judgment set are reselected as the pre-selected lanes and judged.

[0041] Optionally, the acquiring, based on the traffic light information of the road segment on which the target vehicle is currently traveling, an image of each lane of the current road segment and performing analysis and processing includes:

[0042] If the traffic light on the current road section of the target vehicle is a red light for turning left and a green light for going straight, the current green light countdown is obtained and recorded as the light change time;

[0043] Get the straight road image and the left turn plus straight road image of the current road segment;

[0044] The through lane image and the left turn plus through lane image are respectively identified by image recognition technology;

[0045] According to the image recognition results, the number of vehicles in each straight lane and the number of vehicles turning left and going straight are obtained respectively;

[0046] All lanes are entered into the second vehicle judgment set in ascending order according to the number of vehicles in each lane.

[0047] Optionally, the acquiring, based on the traffic light information of the road segment on which the target vehicle is currently traveling, an image of each lane of the current road segment and performing analysis and processing includes:

[0048] The average number of vehicles passing through the green light on the straight road of the current road section per second is obtained through big data, which is recorded as the average passing volume;

[0049] By multiplying the light-changing time by the average passing volume, all vehicles going straight through during the light-changing time are calculated and recorded as the total passing vehicles;

[0050] Select the number of vehicles in the second vehicle judgment set in turn and compare it with the total number of passing vehicles;

[0051] If the number of vehicles in the current element in the second vehicle judgment set is less than the total number of passing vehicles, then the lane type of the current element is determined;

[0052] If the lane type of the current element is a straight road, the current element is recommended as the best road and the lane recommendation strategy ends;

[0053] If the lane type of the current element is a left-turn and straight-through lane, then all vehicles on the left-turn and straight-through lane whose navigation routes are left turns are obtained;

[0054] If the number of vehicles turning left on the above navigation route is not zero, then the current element in the second vehicle judgment set is removed, and the elements in the second vehicle judgment set are reselected for judgment and comparison;

[0055] If the number of vehicles turning left on the above navigation route is zero, the left turn plus the straight road of the current road section is recommended as the best road, and the lane recommendation strategy ends;

[0056] If all elements in the second vehicle judgment set are traversed and are greater than the total number of passing vehicles, the straight road with the smallest number of vehicles corresponding to the elements in the second vehicle judgment set is recommended as the best road, and the lane recommendation strategy ends.

[0057] Optionally, the acquiring, based on the traffic light information of the road segment on which the target vehicle is currently traveling, an image of each lane of the current road segment and performing analysis and processing includes:

[0058] If the traffic light on the current road of the target vehicle is a green light for left turn and a red light for straight ahead, then the straight ahead image and the left turn plus straight ahead image of the current road segment are obtained respectively;

[0059] The through lane image and the left turn plus through lane image are respectively identified by image recognition technology;

[0060] The number of vehicles in each straight lane and the number of vehicles in the left-turn and straight lane are obtained respectively according to image recognition, and each lane is entered into the third vehicle judgment set in ascending order according to the number of vehicles.

[0061] Optionally, the acquiring, based on the traffic light information of the road segment on which the target vehicle is currently traveling, an image of each lane of the current road segment and performing analysis and processing includes:

[0062] The left turn plus the through lane is recorded as the pre-selected lane;

[0063] Obtaining the positions of all vehicles on the pre-selected lane, and numbering the vehicles in ascending order according to the distance between each vehicle on the pre-selected lane and the traffic light of the current road segment, and recording the number in the second left turn set;

[0064] Marking all vehicles whose navigation routes are left turns in the second left turn set in sequence, and entering the number of the marked vehicles in the second left turn set into the number set;

[0065] If the difference between adjacent elements in the numbered set is 1, get the first element in the numbered set;

[0066] If the first element in the numbered set is 1, then get the number of elements in the numbered set;

[0067] Get the countdown of the left turn green light on the current road of the target vehicle, which is recorded as the left turn signal time;

[0068] The average number of vehicles passing through the left-turn and through-the-road green lights per second on the current road section is obtained through big data, and recorded as the average left-turn traffic volume;

[0069] By multiplying the left turn signal time and the average left turn traffic volume, all left-turn vehicles passing through during the left turn signal time are calculated and recorded as the total left-turn traffic volume;

[0070] If the number of elements in the numbered set is greater than the total left-turn traffic, the left-turn plus straight lane in the third vehicle judgment set is eliminated, and the road with the least straight lane vehicles in the third vehicle judgment set is recommended as the best road, and the lane recommendation strategy ends;

[0071] If the number of elements in the numbered set is less than or equal to the total left-turn throughput, the difference between the number of left-turn and through-lane vehicles in the third vehicle judgment set and the number of elements in the numbered set is used as a comparison element;

[0072] By comparing the comparison element with the number of vehicles in each straight lane in the third vehicle judgment set, the lane with the smallest number of vehicles is selected as the best lane, and the lane recommendation strategy is terminated;

[0073] If the first element in the number set is not 1, or the difference between adjacent elements in the number set is not 1, then the left turn and straight lane in the third vehicle judgment set is eliminated;

[0074] The first-ranked element in the third vehicle judgment set is recommended as the best lane, and the lane recommendation strategy ends.

[0075] The present invention has the following beneficial effects:

[0076] 1. This is a traffic image intelligent analysis method based on big data. By determining the road type and the starting and ending points of the road and obtaining all traffic lights on the road, the road is divided into multiple road sections according to the traffic lights and sorted. By sorting the roads in sections and recommending lanes to vehicles in the road sections in turn, it is possible to traverse all road sections in the road from the starting point, and by recommending lane strategies to the recommended vehicles in the road sections in turn, it is possible to unblock traffic flow from the starting point, reasonably allocate road resources, avoid traffic congestion, and improve traffic continuity and overall road traffic efficiency.

[0077] 2. This is a traffic image intelligent analysis method based on big data. It obtains the navigation routes of all vehicles in the road section, screens the vehicles according to the navigation routes, sorts the screened vehicles in the order of passing the traffic lights, and executes the lane recommendation strategy for the sorted vehicles in turn. This can avoid the impact of the lane recommendation strategy on the execution of the lane recommendation strategy between vehicles and ensure the accuracy of the lane recommendation strategy.

[0078] 3. This intelligent traffic image analysis method based on big data can quickly and effectively obtain the number of vehicles in the straight lane and the left turn plus straight lane by identifying and analyzing road images under different traffic light states, which makes the data involved in the lane recommendation strategy more scientific.

[0079] 4. A traffic image intelligent analysis method based on big data. If the traffic lights of the current road section are left-turn red lights and straight-ahead red lights, the number of vehicles in each lane is obtained, and the lane with the least number of vehicles is selected as the pre-selected lane. The type of the pre-selected lane is determined. If it is a straight lane, it is directly recommended. If it is a left turn and a straight lane, the number of vehicles with left turns in all navigation routes within the left turn and the straight lane is obtained. If the number is zero, it is directly recommended. If the number is not zero, the left turn and the straight lane are excluded, and the lane with the least number of vehicles in the lane is re-selected and recommended. By judging the type of the pre-selected lane and determining the strategy for the best lane, unnecessary left-turn lane selection can be avoided. This method improves the traffic efficiency of straight-moving vehicles and helps reduce the situation where the number of left-turning vehicles affects the number of straight-moving vehicles.

[0080] 5. This is a traffic image intelligent analysis method based on big data. If the traffic light of the current road section is a red light for left turn and a green light for straight going, the countdown of the green light for straight going is obtained, and the average traffic volume of the straight road is obtained through big data. The product of the average traffic volume and the green light countdown is calculated, and the difference between the number of vehicles in each straight road in the current road section and the product is used as the actual vehicles on the straight road. By calculating the number of all straight-going vehicles passing within the green light changing time, the lanes can be recommended more scientifically to ensure that the traffic efficiency of straight-going vehicles is maximized within a limited time.

[0081] 6. A traffic image intelligent analysis method based on big data, if the traffic light of the current road section is a left turn green light and a straight red light, by setting the left turn plus straight road as a pre-selected lane, obtaining the left turn green light countdown, obtaining the average left turn traffic volume through big data, calculating the product between the left turn green light countdown and the average left turn traffic volume, and judging whether the number of vehicles whose navigation routes are left turns in the left turn plus straight road is less than the product, if it is less than, obtaining the actual position of the vehicles whose navigation routes are left turns in the left turn plus straight road, if the actual position of all left-turning vehicles is in front of the lane If there is a left-turning vehicle at the front end of the lane and there is no through-going vehicle between adjacent left-turning vehicles, the left-turn plus through lane is recommended. If the actual position of the left-turning vehicle is not at the front end of the lane, or there is a through-going vehicle between adjacent left-turning vehicles, the through lane with the least vehicles in the through lane is recommended. By setting the left-turn plus through lane as the pre-selected road and obtaining the position of the left-turning vehicle, the lane is recommended for the vehicle based on the position of the left-turning vehicle and big data. This can accurately evaluate the passage order and passing capacity of the vehicles, ensure the efficient passage of left-turning and through-going vehicles under complex road conditions, and further improve the overall traffic efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0083] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0084] Example 1, see Figure 1 , a traffic image intelligent analysis method based on big data, comprising:

[0085] Determine the straight road type;

[0086] According to the straight-ahead direction in the straight road, the two ends of the straight road are recorded as the starting point and the ending point;

[0087] Get all traffic lights within a linear road;

[0088] Divide a straight road into multiple road segments according to traffic lights;

[0089] Get traffic light change information;

[0090] Traffic light information includes left turn red light and straight ahead red light, left turn red light and straight ahead green light, left turn green light and straight ahead red light;

[0091] The traffic lights on a straight road only have left turn lights and straight ahead lights. In actual traffic, the traffic light combinations only have three situations: left turn red light and straight ahead red light, left turn red light and straight ahead green light, and left turn green light and straight ahead red light.

[0092] Determine the recommended vehicles and enter the target set;

[0093] The lane recommendation strategy is performed on the elements in the target set in turn. The lane recommendation strategy is as follows:

[0094] Get the lane information of the current road segment, which includes multiple straight lanes and a left turn plus straight lane;

[0095] Obtain the traffic light information of the road section where the target vehicle is currently traveling;

[0096] According to the traffic light information of the road section where the target vehicle is currently traveling, the image of each lane under the current road section is obtained and analyzed and processed;

[0097] According to the image analysis processing result of each lane in the current road section, the vehicle information of each lane in the current road section is determined;

[0098] Determine the best lane that matches the recommended vehicle navigation route through the vehicle information and traffic light information of each lane of the current road segment;

[0099] By determining the road type and the starting and ending points of the road, and obtaining all traffic lights on the road, the road is divided into multiple road sections according to the traffic lights and sorted. By sorting the road in sections and recommending lanes to vehicles in the road sections in turn, it is possible to traverse all road sections in the road from the starting point, and implement lane recommendation strategies for recommended vehicles in the road sections in turn. This way, traffic flow can be unblocked from the starting point, road resources can be reasonably allocated, traffic congestion can be avoided, and traffic continuity and overall road traffic efficiency can be improved.

[0100] The step of determining the recommended vehicle and entering the target vehicle set includes:

[0101] Sort all road segments from nearest to farthest from the starting point;

[0102] The first ranked road segment is taken as the target road segment;

[0103] Obtain the navigation routes of all vehicles in the target road segment;

[0104] The vehicles whose navigation routes are to pass the current traffic light and go straight are taken as recommended vehicles and are entered into the target vehicle set;

[0105] Sort all elements in the target vehicle set according to the order in which they pass the traffic lights of the current road segment;

[0106] Select the elements in the target vehicle set as the target vehicle in sequence;

[0107] By obtaining the navigation routes of all vehicles in the road section and screening the vehicles according to the navigation routes, the screened vehicles are sorted in the order of passing the traffic lights, and the lane recommendation strategy is executed on the sorted vehicles in turn. This can avoid the impact of the lane recommendation strategy on other vehicles and ensure the accuracy of the lane recommendation strategy.

[0108] The method of obtaining an image of each lane of the current road segment according to the traffic light information of the road segment on which the target vehicle is currently traveling and performing analysis and processing includes:

[0109] Get the traffic light of the current road section of the target vehicle;

[0110] If the traffic lights of the current road section of the target vehicle are a left-turn red light and a straight-through red light, then the straight-through image and the left-turn plus straight-through image of the current road section are obtained;

[0111] The through lane image and the left turn plus through lane image are respectively identified by image recognition technology;

[0112] By identifying and analyzing road images under different traffic light conditions, the number of vehicles in the straight lane and the left-turn and straight lane can be quickly and effectively obtained, which makes the data involved in the lane recommendation strategy more scientific.

[0113] According to the image recognition results, the number of vehicles in each straight lane and the number of vehicles turning left and going straight are obtained respectively;

[0114] All lanes are entered into the first vehicle judgment set in ascending order according to the number of vehicles in each lane.

[0115] The determining of the best lane that meets the recommended vehicle navigation route through the vehicle information and traffic light information of each lane of the current road segment includes:

[0116] Select the elements in the first vehicle judgment set in sequence and record them as pre-selected lanes;

[0117] If the pre-selected lane is a straight road, the pre-selected lane is recommended as the best road and the lane recommendation strategy ends;

[0118] If the pre-selected lane is a left-turn and straight lane, the navigation routes of all vehicles in the left-turn and straight lane are obtained;

[0119] Mark all vehicles whose navigation routes are left turns and enter the first left turn set;

[0120] If the number of elements in the left-turn set is zero, then the left turn plus the straight road is recommended as the best road, and the lane recommendation strategy ends;

[0121] If the number of elements in the left-turn set is not zero, the left-turn plus straight lane in the first vehicle judgment set is removed, and the elements in the first vehicle judgment set are reselected as the pre-selected lanes and judged;

[0122] By obtaining the number of vehicles in each lane, the lane with the least number of vehicles is selected as the pre-selected lane, and the type of the pre-selected lane is determined. If it is a straight lane, it is directly recommended. If it is a left turn and a straight lane, the number of vehicles with left turns in all navigation routes within the left turn and the straight lane is obtained. If the number is zero, it is directly recommended. If the number is not zero, the left turn and the straight lane are excluded, and the lane with the least number of vehicles in the lane is re-selected and recommended. By determining the type of the pre-selected lane and determining the strategy for the best lane, unnecessary left-turn lane selection can be avoided. This method improves the traffic efficiency of straight-moving vehicles and helps reduce the situation where the number of left-turning vehicles affects the number of straight-moving vehicles.

[0123] The method of obtaining an image of each lane of the current road segment according to the traffic light information of the road segment on which the target vehicle is currently traveling and performing analysis and processing includes:

[0124] If the traffic light on the current road section of the target vehicle is a red light for turning left and a green light for going straight, the current green light countdown is obtained and recorded as the light change time;

[0125] Get the straight road image and the left turn plus straight road image of the current road segment;

[0126] The through lane image and the left turn plus through lane image are respectively identified by image recognition technology;

[0127] According to the image recognition results, the number of vehicles in each straight lane and the number of vehicles turning left and going straight are obtained respectively;

[0128] All lanes are entered into the second vehicle judgment set in ascending order according to the number of vehicles in each lane.

[0129] The method of obtaining an image of each lane of the current road segment according to the traffic light information of the road segment on which the target vehicle is currently traveling and performing analysis and processing includes:

[0130] The average number of vehicles passing through the green light on the straight road of the current road section per second is obtained through big data, which is recorded as the average passing volume;

[0131] By multiplying the light-changing time by the average passing volume, all vehicles going straight through during the light-changing time are calculated and recorded as the total passing vehicles;

[0132] Select the number of vehicles in the second vehicle judgment set in turn and compare it with the total number of passing vehicles;

[0133] If the number of vehicles in the current element in the second vehicle judgment set is less than the total number of passing vehicles, then the lane type of the current element is determined;

[0134] If the lane type of the current element is a straight road, the current element is recommended as the best road and the lane recommendation strategy ends;

[0135] If the lane type of the current element is a left-turn and straight-through lane, then all vehicles on the left-turn and straight-through lane whose navigation routes are left turns are obtained;

[0136] If the number of vehicles turning left on the above navigation route is not zero, then the current element in the second vehicle judgment set is removed, and the elements in the second vehicle judgment set are reselected for judgment and comparison;

[0137] If the number of vehicles turning left on the above navigation route is zero, the left turn plus the straight road of the current road section is recommended as the best road, and the lane recommendation strategy ends;

[0138] If all elements in the second vehicle judgment set are traversed and are greater than the total number of passing vehicles, the straight road with the smallest number of vehicles corresponding to the elements in the second vehicle judgment set is recommended as the best road, and the lane recommendation strategy ends;

[0139] Get the countdown of the green light for going straight, get the average traffic volume of the straight road through big data, calculate the product of the average traffic volume and the green light countdown, and take the difference between the number of vehicles on each straight road in the current road section and the product as the actual number of vehicles on the straight road. By calculating the number of all straight-going vehicles passing within the green light changing time, lanes can be recommended more scientifically to ensure that the traffic efficiency of straight vehicles is maximized within a limited time.

[0140] The method of obtaining an image of each lane of the current road segment according to the traffic light information of the road segment on which the target vehicle is currently traveling and performing analysis and processing includes:

[0141] If the traffic light on the current road of the target vehicle is a green light for left turn and a red light for straight ahead, then the straight ahead image and the left turn plus straight ahead image of the current road segment are obtained respectively;

[0142] The through lane image and the left turn plus through lane image are respectively identified by image recognition technology;

[0143] The number of vehicles in each straight lane and the number of vehicles in the left-turn and straight lane are obtained respectively according to image recognition, and each lane is entered into the third vehicle judgment set in ascending order according to the number of vehicles.

[0144] The method of obtaining an image of each lane of the current road segment according to the traffic light information of the road segment on which the target vehicle is currently traveling and performing analysis and processing includes:

[0145] The left turn plus the through lane is recorded as the pre-selected lane;

[0146] Obtaining the positions of all vehicles on the pre-selected lane, and numbering the vehicles in ascending order according to the distance between each vehicle on the pre-selected lane and the traffic light of the current road segment, and recording the number in the second left turn set;

[0147] Marking all vehicles whose navigation routes are left turns in the second left turn set in sequence, and entering the number of the marked vehicles in the second left turn set into the number set;

[0148] If the difference between adjacent elements in the numbered set is 1, get the first element in the numbered set;

[0149] If the first element in the numbered set is 1, then get the number of elements in the numbered set;

[0150] Get the countdown of the left turn green light on the current road of the target vehicle, which is recorded as the left turn signal time;

[0151] The average number of vehicles passing through the left-turn and through-the-road green lights per second on the current road section is obtained through big data, and recorded as the average left-turn traffic volume;

[0152] By multiplying the left turn signal time and the average left turn traffic volume, all left-turn vehicles passing through during the left turn signal time are calculated and recorded as the total left-turn traffic volume;

[0153] If the number of elements in the numbered set is greater than the total left-turn traffic, the left-turn plus straight lane in the third vehicle judgment set is eliminated, and the road with the least straight lane vehicles in the third vehicle judgment set is recommended as the best road, and the lane recommendation strategy ends;

[0154] If the number of elements in the numbered set is less than or equal to the total left-turn throughput, the difference between the number of left-turn and through-lane vehicles in the third vehicle judgment set and the number of elements in the numbered set is used as a comparison element;

[0155] By comparing the comparison element with the number of vehicles in each straight lane in the third vehicle judgment set, the lane with the smallest number of vehicles is selected as the best lane, and the lane recommendation strategy is terminated;

[0156] If the first element in the number set is not 1, or the difference between adjacent elements in the number set is not 1, then the left turn and straight lane in the third vehicle judgment set is eliminated;

[0157] Recommend the first ranked element in the third vehicle judgment set as the best lane, and end the lane recommendation strategy;

[0158] By setting the left turn and straight lane as the pre-selected lane, obtaining the left turn green light countdown, obtaining the average left turn throughput through big data, calculating the product between the left turn green light countdown and the average left turn throughput, judging whether the number of vehicles with left turns in the left turn and straight lane is less than the product, if it is less than, obtaining the actual position of the vehicles with left turns in the left turn and straight lane. If the actual positions of all left-turning vehicles are at the front end of the lane and there are no straight-moving vehicles between adjacent left-turning vehicles, the left turn and straight lane is recommended. If the actual position of a left-turning vehicle is not at the front end of the lane, or there is a straight-moving vehicle between adjacent left-turning vehicles, the straight lane with the least vehicles in the straight lane is recommended. By setting the left turn and straight lane as the pre-selected road and obtaining the position of the left-turning vehicle, the lane is recommended for the vehicle according to the position of the left-turning vehicle and the big data. This can accurately evaluate the traffic order and passing capacity of the vehicles, ensure the efficient passage of left-turning and straight-moving vehicles under complex road conditions, and further improve the overall traffic efficiency.

[0159] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0160] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A traffic image intelligent analysis method based on big data, characterized by: include: Determine the straight road type; According to the straight-ahead direction in the straight road, the two ends of the straight road are recorded as the starting point and the ending point; Get all traffic lights within a linear road; Divide a straight road into multiple road segments according to traffic lights; Get traffic light change information; Traffic light information includes left turn red light and straight ahead red light, left turn red light and straight ahead green light, left turn green light and straight ahead red light; Determine the recommended vehicles and enter the target set; The lane recommendation strategy is performed on the elements in the target set in turn. The lane recommendation strategy is as follows: Get the lane information of the current road segment, which includes multiple straight lanes and a left turn plus straight lane; Obtain the traffic light information of the road section where the target vehicle is currently traveling; According to the traffic light information of the road section where the target vehicle is currently traveling, the image of each lane under the current road section is obtained and analyzed and processed; According to the image analysis processing result of each lane in the current road section, the vehicle information of each lane in the current road section is determined; The best lane that matches the recommended vehicle navigation route is determined through the vehicle information and traffic light information of each lane of the current road segment.

2. The method for intelligent analysis of traffic images based on big data according to claim 1 is characterized by: The step of determining the recommended vehicle and entering the target vehicle set includes: Sort all road segments from nearest to farthest from the starting point; The first ranked road segment is taken as the target road segment; Obtain the navigation routes of all vehicles in the target road segment; The vehicles whose navigation routes are to pass the current traffic light and go straight are taken as recommended vehicles and are entered into the target vehicle set; Sort all elements in the target vehicle set according to the order in which they pass the traffic lights of the current road segment; Select the elements in the target vehicle set as the target vehicle in sequence.

3. The method for intelligent analysis of traffic images based on big data according to claim 1 is characterized by: The method of acquiring an image of each lane of the current road segment according to the traffic light information of the road segment on which the target vehicle is currently traveling and performing analysis and processing includes: Get the traffic light of the current road section of the target vehicle; If the traffic lights of the current road section of the target vehicle are a left-turn red light and a straight-through red light, then the straight-through image and the left-turn plus straight-through image of the current road section are obtained; The through lane image and the left turn plus through lane image are respectively identified by image recognition technology; According to the image recognition results, the number of vehicles in each straight lane and the number of vehicles turning left and going straight are obtained respectively; All lanes are entered into the first vehicle judgment set in ascending order according to the number of vehicles in each lane.

4. The method for intelligent analysis of traffic images based on big data according to claim 1, characterized in that: The determining of the best lane that meets the recommended vehicle navigation route based on the vehicle information and traffic light information of each lane of the current road segment includes: Select the elements in the first vehicle judgment set in sequence and record them as pre-selected lanes; If the pre-selected lane is a straight road, the pre-selected lane is recommended as the best road and the lane recommendation strategy ends; If the pre-selected lane is a left-turn and straight lane, the navigation routes of all vehicles in the left-turn and straight lane are obtained; Mark all vehicles whose navigation routes are left turns and enter the first left turn set; If the number of elements in the left-turn set is zero, then the left turn plus the straight road is recommended as the best road, and the lane recommendation strategy ends; If the elements in the left-turn set are not zero, the left-turn plus straight lane in the first vehicle judgment set is eliminated, and the elements in the first vehicle judgment set are reselected as the pre-selected lanes and judged.

5. The method for intelligent analysis of traffic images based on big data according to claim 1 is characterized by: The method of acquiring an image of each lane of the current road segment according to the traffic light information of the road segment on which the target vehicle is currently traveling and performing analysis and processing includes: If the traffic light on the current road section of the target vehicle is a red light for turning left and a green light for going straight, the current green light countdown is obtained and recorded as the light change time; Get the straight road image and the left turn plus straight road image of the current road segment; The through lane image and the left turn plus through lane image are respectively identified by image recognition technology; According to the image recognition results, the number of vehicles in each straight lane and the number of vehicles turning left and going straight are obtained respectively; All lanes are entered into the second vehicle judgment set in ascending order according to the number of vehicles in each lane.

6. The method for intelligent analysis of traffic images based on big data according to claim 1 is characterized by: The method of acquiring an image of each lane of the current road segment according to the traffic light information of the road segment on which the target vehicle is currently traveling and performing analysis and processing includes: The average number of vehicles passing through the green light on the straight road of the current road section per second is obtained through big data, which is recorded as the average passing volume; By multiplying the light-changing time by the average passing volume, all vehicles going straight through during the light-changing time are calculated and recorded as the total passing vehicles; Select the number of vehicles in the second vehicle judgment set in turn and compare it with the total number of passing vehicles; If the number of vehicles in the current element in the second vehicle judgment set is less than the total number of passing vehicles, then the lane type of the current element is determined; If the lane type of the current element is a straight road, the current element is recommended as the best road and the lane recommendation strategy ends; If the lane type of the current element is a left-turn and straight-through lane, then all vehicles on the left-turn and straight-through lane whose navigation routes are left turns are obtained; If the number of vehicles turning left on the above navigation route is not zero, then the current element in the second vehicle judgment set is removed, and the elements in the second vehicle judgment set are reselected for judgment and comparison; If the number of vehicles turning left on the above navigation route is zero, the left turn plus the straight road of the current road section is recommended as the best road, and the lane recommendation strategy ends; If all elements in the second vehicle judgment set are traversed and are greater than the total number of passing vehicles, the straight road with the smallest number of vehicles corresponding to the elements in the second vehicle judgment set is recommended as the best road, and the lane recommendation strategy ends.

7. The method for intelligent analysis of traffic images based on big data according to claim 1 is characterized by: The method of acquiring an image of each lane of the current road segment according to the traffic light information of the road segment on which the target vehicle is currently traveling and performing analysis and processing includes: If the traffic light on the current road of the target vehicle is a green light for left turn and a red light for straight ahead, then the straight ahead image and the left turn plus straight ahead image of the current road segment are obtained respectively; The through lane image and the left turn plus through lane image are respectively identified by image recognition technology; The number of vehicles in each straight lane and the number of vehicles in the left-turn and straight lane are obtained respectively according to image recognition, and each lane is entered into the third vehicle judgment set in ascending order according to the number of vehicles.

8. The method for intelligent analysis of traffic images based on big data according to claim 1 is characterized by: The method of acquiring an image of each lane of the current road segment according to the traffic light information of the road segment on which the target vehicle is currently traveling and performing analysis and processing includes: The left turn plus the through lane is recorded as the pre-selected lane; Obtaining the positions of all vehicles on the pre-selected lane, and numbering the vehicles in ascending order according to the distance between each vehicle on the pre-selected lane and the traffic light of the current road segment, and recording the number in the second left turn set; Marking all vehicles whose navigation routes are left turns in the second left turn set in sequence, and entering the number of the marked vehicles in the second left turn set into the number set; If the difference between adjacent elements in the numbered set is 1, get the first element in the numbered set; If the first element in the numbered set is 1, then get the number of elements in the numbered set; Get the countdown of the left turn green light on the current road of the target vehicle, which is recorded as the left turn signal time; The average number of vehicles passing through the left-turn and through-the-road green lights per second on the current road section is obtained through big data, and recorded as the average left-turn traffic volume; By multiplying the left turn signal time and the average left turn traffic volume, all left-turn vehicles passing through during the left turn signal time are calculated and recorded as the total left-turn traffic volume; If the number of elements in the numbered set is greater than the total left-turn traffic, the left-turn plus straight lane in the third vehicle judgment set is eliminated, and the road with the least straight lane vehicles in the third vehicle judgment set is recommended as the best road, and the lane recommendation strategy ends; If the number of elements in the numbered set is less than or equal to the total left-turn throughput, the difference between the number of left-turn and through-lane vehicles in the third vehicle judgment set and the number of elements in the numbered set is used as a comparison element; By comparing the comparison element with the number of vehicles in each straight lane in the third vehicle judgment set, the lane with the smallest number of vehicles is selected as the best lane, and the lane recommendation strategy is terminated; If the first element in the number set is not 1, or the difference between adjacent elements in the number set is not 1, then the left turn and straight lane in the third vehicle judgment set is eliminated; The first-ranked element in the third vehicle judgment set is recommended as the best lane, and the lane recommendation strategy ends.

Images